INDEX
Explanations
proper nouns, especially place names and brand names, along with some isolated words in Arabic
New Auto-Interp
Negative Logits
<bos>
-1.02
for
-0.70
to
-0.67
when
-0.65
but
-0.64
and
-0.62
it
-0.62
in
-0.58
↵
-0.57
at
-0.56
POSITIVE LOGITS
يتيمه
1.38
تانيه
1.08
MainAxisSize
1.07
Efq
1.06
AndEndTag
1.03
بوابة
1.03
متعلقه
1.00
Theſe
0.99
tvguidetime
0.94
aarrggbb
0.94
Activations Density 1.161%